Methods

Every method the add-in can rank has a short tutorial: what it assumes, a concrete sample question, the numbers, step-by-step on those numbers, and a chart of the same series. Ranking in Excel is still rolling-origin MASE — “best supported among candidates,” never the true model.

Free analysis never uploads your series. These pages use canned teaching numbers produced by the same forecast engine as the add-in.

Baselines

Naive

Repeat the last observation. The baseline every other method must beat.

Seasonal naive

Copy last year’s value for the same season.

Drift

Random walk plus a constant average change.

Smoothing

Moving average

Average the last k points; the window is chosen by fit among {3, 5, 7, 12}.

Simple exponential smoothing

A fading memory of the local level — no trend, no season.

Holt linear

Local level plus a linear trend (α and β).

Damped trend

A trend that flattens (φ < 1) instead of running forever.

Holt–Winters additive

Level + trend + a seasonal wave of constant width.

Holt–Winters multiplicative

Seasonal swings grow with the level; requires y > 0.

Intermittent

Croston

Split size and interval; update only when demand occurs.

Syntetos–Boylan

Croston with a (1 − α/2) bias correction — usual spare-parts default.

Teunter–Syntetos–Babai

Demand probability updates every period, so a dying item can decay to zero.

Regression

Time-series linear model

OLS on trend plus Fourier seasonal terms, and optional regressors.

Decomposition

Decompose + seasonal naive

Pull off a seasonal pattern, then naive the remainder.

ARIMA

ARIMA

Bounded (p,d,q ≤ 2) ARMA after a few differences, picked by AICc.

SARIMA

ARIMA plus a seasonal-difference proxy; needs n ≥ 3m.

Theta

Theta

SES plus a drift line — the M3-famous combination (Pro).

Piecewise + Fourier

Piecewise trend + Fourier

One midpoint changepoint plus Fourier season — in-engine, no Python Prophet.

Volatility

GARCH(1,1)

Variance clustering on returns — not a price oracle. Educational, not advice.

TBATS

TBATS-lite

Trigonometric seasonality stand-in for multiple seasonalities (Pro).

Ensemble

Ensemble mean

Average top-k Free point forecasts when several methods tie on MASE (Pro).

ETS

ETS auto

Pick the best of SES / Holt / damped / Holt–Winters on SSE, then backtest (Pro).

Educational only. Not investment, weather, or operational advice. In the add-in, rank is rolling-origin MASE — “best supported among candidates on this series.”